http://dspace.cityu.edu.hk:80/handle/2031/708 2013-05-01T03:06:35Z http://dspace.cityu.edu.hk:80/handle/2031/6603 Title: Incentive ratio and market equilibrium Authors: Zhang, Jie ( 張捷) Abstract: The thesis considers Fisher market with Constant Elasticity of Substitution (CES) utilities, including Linear utility, Leontief utility and Cobb-Douglas utility, and studies strategic behaviors of individual buyers in market equilibria. The notion of incentive ratio is introduced to capture the extent to which utilities can be increased by strategic behaviors of individuals. Formally, in a given market, the incentive ratio of any given buyer is defined to be his maximum possible utility obtainable by employing strategic behaviors divided by his utility when bidding truthfully, given any fixed bids of all other buyers. The incentive ratio of the whole market is defined to be the maximum incentive ratio of all buyers. Note that the ratio is at least one which happens for an incentive compatible mechanism. While buyers may get larger utilities when they behave strategically, it is interesting to know the range of its improvement. This thesis shows that the benefits of strategic behaviors may be quite limited. For Leontief market, the incentive ratio is at most 2, and we give an example to show that this upper bound is tight. Moreover, the incentive ratio of Leontief market is insensitive to market size. For Linear market, the thesis presents a constructive proof for the incentive ratio of 2-buyer market. For Cobb-Douglas market, the thesis derive the best response strategy of the buyers and the su±cient and necessary condition for bidding truthful to be a dominant strategy for all buyers. We also prove the incentive ratio is independent of the number of buyers. Meanwhile, simulation results show that the incentive ratio of Cobb-Douglas market is indeed small. In addition, the thesis obtains a matching exponential algorithmic bound for TUCKER in the oracle model, in all dimensions. The lower bound is based on a reduction from the Direction Preserving Zero Point (DPZP) to TUCKER. DPZP is a new mathematical structure introduced recently as an appropriate technical tool to formulate the more traditional fixed point computational problems. The upper bound requires a combinatorial parity lemma that relates the existence of the complementary edges on the boundary with certain parity properties of the fully-colored simplices on the boundary of the hypergrid. The connection also allows us to derive an immediate proof that TUCKER is PPAD-complete for all constant dimensions. Notes: CityU Call Number: HB145 .Z428 2011; vii, 106 leaves 30 cm.; Thesis (Ph.D.)--City University of Hong Kong, 2011.; Includes bibliographical references (leaves 97-106) 2011-01-01T00:00:00Z http://dspace.cityu.edu.hk:80/handle/2031/6602 Title: Smartphone security in ubiquitous networking environments Authors: Xu, Nan ( 徐楠) Abstract: Currently researchers mainly focus on new countermeasures against malware, e.g. new malware analysis and detection technologies. With structure and functionality similar to computers, Smartphones also face the threat of the attacks of mobile malware. In terms of personal communication platform, the big market of Smartphones attracts and leaves room for malicious hackers to exploit the vulnerabilities in Smartphones for mischiefs or, even worse, for criminal gains. In terms of mobile social network platform, the ubiquitous networking security vulnerabilities have spread from the Internet to 3G, WiFi, and LTE networks. There is a chain of communication traversing multiple networks, enabling hackers to attack the weakest link - Smartphones, and hence causing significant damages to otherwise well protected networks. In this light, researchers gradually turn to focus on the Smartphones security. In this thesis, we make our contributions from following three aspects. Stealthy Video Capturer - New Video-based Spyware in 3G Smartphones As personal communication platforms, 3G Smartphones have become increasingly pervasive in recent years. Unfortunately they contain much personal and confidential information and are easily compromised. We will first address the privacyrelated Smartphone security, as well as relevant typical applications. We designed and implemented Stealthy Video Capturer (SVC), a new video-based spyware, on real 3G Smartphones. SVC can steal users' private information through secretly recording videos for the third party and transmitting them to that party. To our knowledge, our work is the first one that investigates video-based vulnerabilities in 3G Smartphones. We illustrate two practical applications of SVC, i.e. mobile forensics and child care. We anticipate that our work will draw widespread attention to this issue. 3G Real Time Video-based Mobile Guarder We investigate a novel technique to acquire comprehensive context information of remote Smartphone users, including surrounding environment, location and behaviors, for further information process, e.g. localization and human safeguard. Our system is simply launched by a normal 3G video call and to transmit the captured real time video in 3G connections. The technique makes use of the efficiency of 3G video call and thus zero-configuration for the call initiator. It can achieve the remote user's status on-demand and fast show a general view of the user for the call initiator, besides location based service. Moreover, no costly special devices other than normal 3G Smartphones are needed in our technique. It can also help to revise the inherent limitations of GPS by employing the surrounding environment information provided by captured real time video when lies indoor or dense high building districts. Smartphone Operating System Fingerprinting As mobile social network platform, there is an unavoidable potential threat for infected Smartphones propagating mobile malware to others in the ubiquitous networking environment. Moreover, stimulated by the fast growing popularity of wireless-enabled Smartphones and the inherent vulnerability of Smartphone operating system (OS), we investigate a new fingerprinting technique that identifies the Smartphone OS. The technique is greatly helpful for wireless network administrators to adaptively accommodate the network security policy based on the real time fingerprinting data. Different from existing schemes on Personal Computer (PC) OS fingerprinting, our fingerprinting technique is launched during the WiFi connection period of a Smartphone. The technique makes use of a new reliable metric called channel dwell time, which is the Smartphone's stay time on a certain channel in the normal probing process of Smartphones. Our technique can quickly identify the Smartphone operating system with high accuracy based on channel dwell time without any modification to or specific interaction with the Smartphone. Moreover no costly special apparatuses other than several normal PC or notebooks installed with WiFi cards are needed. Then we implement the technique and conduct extensive experiments with real world popular Smartphone OSes. Our experimental evaluation shows that our technique is applicable to most commercial Smartphones and reaches an average identification accuracy of up to 98.1%. This thesis is organized as follows. Chapter 1 describes the background of the thesis. Chapter 2 introduces the related work. Chapter 3 investigates the video-based vulnerabilities in 3G Smartphones, especially a new video-based spyware on 3G Smartphones. Chapter 4 develops an novel technique to acquire comprehensive context information of remote Smartphone users for further information process, e.g. localization and human safeguard. Chapter 5 exploits a new reliable metric called channel dwell time to quickly identify Smartphone OSes. Chapter 6 concludes this thesis. Notes: CityU Call Number: TK5103.485 .X84 2010; vii, 112 leaves : ill. 30 cm.; Thesis (Ph.D.)--City University of Hong Kong, 2010.; Includes bibliographical references (leaves 105-112) 2010-01-01T00:00:00Z http://dspace.cityu.edu.hk:80/handle/2031/6601 Title: QoS multicast routing and causal ordered group communications in cognitive radio networks Authors: Xie, Liming ( 謝黎明) Abstract: With the development of wireless communications, wireless networks become more and more popular in our lives. However, there exist some problems in the usage of spectrum. On one hand, there has not much unlicensed spectrum left. On the other hand, the utilization of the licensed spectrum is very low. Such problems seriously block the further development of wireless communications. Fortunately, cognitive radio technology can reduce the unbalance of the spectrum usage. That is because cognitive radio is able to capture or sense the idle spectrum information from its radio environment and accesses such idle spectrum without interfering with the licensed user. Cognitive radio networks have received a lot of attentions in recent years because they can exploit the existing wireless spectrum opportunistically and improve the spectral efficiency. In such networks, the secondary users equipped with cognitive radio can sense and access the "spectrum hole" unoccupied by the primary user. Since nodes have different surrounding environment in cognitive radio networks, the secondary users have different available spectrum bands or channels. This heterogeneous characteristic introduces additional complexities to communications in cognitive radio networks. Based on the heterogeneous characteristic, we address the issue of QoS multicast routing and the issue of supporting causal ordered group communication in cognitive radio networks. The major work and contributions can be summarized as follows. (1) Discuss the issues of QoS multicast routing and transmission scheduling in cognitive radio ad hoc networks. The problem of our concern is: given a cognitive radio network and a QoS multicast request, how to set up a multicast connection so that the total bandwidth consumption of the multicast is minimized while the QoS requirements are met. The QoS requirements contain the delay and bandwidth in this thesis. In order to solve the problem, we propose a four-way handshake and distributed protocol which consists of tree construction and transmission scheduling. To minimize the bandwidth consumption, we introduce an auxiliary graph to represent timeslot assignment in the network. The transmission scheduling problem is then transformed into the minimal set cover problem and a heuristic method is proposed to it. Extensive simulations are conducted to show that our proposed method performs well in reducing the total bandwidth consumption of the multicast tree. Additionally, our method has high success rate. (2) Further discuss the issues of QoS multicast routing in cognitive radio ad hoc networks and propose two methods to solve it. One is a two-phase method. In this method, we first employ a minimal spanning tree based algorithm to construct a multicast tree, and then use a transmission scheduling algorithm to reserve timeslots for each non leaf nodes such that the bandwidth consumption of the tree is minimized. The other is an integrated method that considers the multicast routing together with slot reservation. It starts with the destinations which are considered as the subtrees, and merges the subtrees to a single tree towards the direction of the source. Extensive simulations are conducted to show that the two methods perform much better than the shortest path tree based two-phase method in reducing the total bandwidth consumption of the multicast tree. Besides, the two methods have higher success rate. (3) Discuss the causal ordering group communication in cognitive radio ad hoc networks. Different from the previous work, we discuss the causal ordering at the network layer. The problem of our concern is: given a group communication request in a cognitive radio ad hoc network, how to set up the connection so that the causal ordering of the group communication can be preserved, and the bandwidth consumption of the communication is minimized. We propose a non-core group routing method to solve it. In the method, we first construct a multicast tree for the communication, and then assign slots for all tree links. It is proved that the method can preserve the causal ordering of messages without extra communication overhead nor the latency for delivering messages. Simulations are conducted to show that compared with the core based tree based method, our method requires less bandwidth consumption of the group communication, and has higher success rate. Notes: CityU Call Number: TK5103.4815 .X53 2011; vii, 103 leaves : ill. 30 cm.; Thesis (Ph.D.)--City University of Hong Kong, 2011.; Includes bibliographical references (leaves [93]-103) 2011-01-01T00:00:00Z http://dspace.cityu.edu.hk:80/handle/2031/6600 Title: Optimizations of power allocation Authors: Wu, Weiwei ( 吴巍煒) Abstract: The number of mobile users (devices) reaches 5 billion in year 2010. Due to the great number of devices and the fact that all the devices are equipped with limited battery (power), designing algorithms to save the energy usage on a single device or properly utilize the power to maximize the date throughput of the mobile system receives many concerns. We theoretically study several scheduling/allocation problems to optimize the power usage on a device and the system. We provide a thorough study on approximation algorithms (or exact) to compute the (optimal) power allocation on a single device and online algorithms with optimal competitive ratio to maximize the throughput on a mobile system. Dynamic voltage scaling technique provides the capability for processors to adjust the speed and control the energy consumption. We study the pessimistic accelerate model where the acceleration rate of the processor speed is at most K and jobs cannot be executed during the speed transition period. The objective is to find a min‐energy (optimal) schedule that finishes every job within its deadline. The aligned jobs where earlier released jobs have earlier deadlines are studied. We start by investigating a special case where all jobs have a common arrival time and design an O(n2) algorithm to compute the optimal schedule based on some nice properties of the optimal schedule. Then, we study the general aligned jobs and obtain an O(n2) algorithm to compute the optimal schedule by using the algorithm for the common arrival time case as a building block. Because our algorithm relies on the computation of the optimal schedule in the ideal model (K= ∞), in order to achieve O(n2) complexity, we improve the complexity of computing the optimal schedule in the ideal model for aligned jobs from the currently best known O(n2logn) to O(n2). We then turn to the power allocation problem on the multi‐machine system (or devices on the mobile networks). The limited power, dynamic status and interferences are all serious concerns of the wireless channels. Apart from considering the physical characteristics, another important direction is to study the selfish behaviors of the users (e.g. transmitters). We study the online power allocation problem on the time varying channels to maximize the global gains of the network with interference. Besides the physical constraints, by treating each user as an individual agent (player) in a game theoretical view, we simultaneously capture the selfish behavior of the users. The users are allowed to cheat on their private value (power budget) to increase their own utility. We aim at maximizing the global gains of the network while ensuring that each agent optimizes its own utility by reporting its true power budget (admits truthful mechanism). Let the channel quality of all users be varying over time in range [h, H] and it equals 0 when that user is not allowed to transmit any data in that time. We derive a general lower bound Ω(log H/h) - competitive for any randomized algorithm using Yao's minimax principle. We further propose a randomized online algorithm which not only admits truthful mechanism but also is proved O(log H/h) - competitive in expectation and hence optimal with respect to the asymptotic competitive ratio. Notes: CityU Call Number: TK7895.P68 W8 2011; 99 leaves 30 cm.; Thesis (Ph.D.)--City University of Hong Kong, 2011.; Includes bibliographical references (leaves 96-99) 2011-01-01T00:00:00Z